Detecting Security Vulnerabilities in Web Applications Using Dynamic Analysis with Penetration Testing Andrey Petukhov, Dmitry Kozlov Computing Systems Lab, Department of Computer Science, Moscow State University
[email protected],
[email protected] Abstract. The number of reported web application vulnerabilities is increasing dramatically. The most of vulnerabilities result from improper input validation. This paper presents extensions to the Tainted Mode model which allows inter- module vulnerabilities detection. Besides, this paper presents a new approach to vulnerability analysis which incorporates advantages of penetration testing and dynamic analysis. This approach effectively utilizes the extended Tainted Mode model. Keywords: Web Applications, Second-order injection, Vulnerability Analysis, Penetration Testing, Dynamic Analysis, Taint Analysis. 1 Introduction Security vulnerabilities in web applications may result in stealing of confidential data, breaking of data integrity or affect web application availability. Thus the task of se- curing web applications is one of the most urgent for now: according to Acunetix sur- vey [1] 60% of found vulnerabilities affect web applications. The most common way of securing web applications is searching and eliminating vulnerabilities therein. Ex- amples of another ways of securing web application include safe development [38-40], implementing intrusion detection and/or protection systems [32, 35-36], and web application firewalls [37]. According to OWASP [8], the most efficient way of finding security vulnerabilities in web applications is manual code review. This technique is very time-consuming, requires expert skills, and is prone to overlooked errors. Therefore, security society actively develops automated approaches to finding security vulnerabilities. These ap- proaches can be divided into two wide categories: black-box and white-box testing.